People detection

[SteveMann]

I am working on a project for Halloween that has many parts. One of my gags is a skull or a whole skeleton with an 8x8 bi-color led matrix for each eye. I spent a few days working on face detection and the eyes "follow" the person as long as the camera inside the skull can see the face. It's slow, but pretty realistic. (It freaked my wife out- and she knows what I am working on). In the future I plan to turn the head with a stepper motor if the face goes to the edge of the image.

But my wife noticed something basically wrong with my face detection. The guests will likely be wearing a mask.

(darn).

So, can anyone think of another way to detect where a person is located. It doesn't have to be precise. I was thinking of using a pair of microphones into an A to D and move the eyes toward the louder of the two. I though of using a couple of PIR sensors and turn toward the first that sees movement. Maybe an x-y matrix on the floor to detect where someone is standing?

Any ideas would be appreciated.

[sparkyhall]

You may be able to use the Pi camera motions vectors like PiKrellCam does to detect motion:

http://billw2.github.io/pikrellcam/pikrellcam.html

[daveyman123]

https://www.youtube.com/watch?v=oJAl9Yd3kNo
this guy manages to follow people with a laser

TBH it looks like youre in for a lot more tinkering with opencv as opencv seems to be the best.. although it may also be dark on halloween. so idk

sentdex has some tutorials on opencv tracking.

[SteveMann]

You may be able to use the Pi camera motions vectors like PiKrellCam does to detect motion:

http://billw2.github.io/pikrellcam/pikrellcam.html

Thanks for the tip. I don't really need to detect a face, just where people may be. The goal is to have the eyes "follow" people on the porch.